As part of the HetNet mobility enhancements for LTE, one of the topics being discussed in RAN2 is the accuracy of mobility state estimation (MSE) mechanism. I came across a nice description of the existing MSE scheme in a RAN2 doc (R2-115919). The 3GPP TS36.331 (RRC specification) describes MSE for UEs in RRC_CONNECTED mode as follows:
Step 1: The UE counts handovers during time period t-Evaluation and t-HystNormal.
Note: The UE shall not count consecutive handovers between same two cells.
Step 2: The UE detects its own mobility state based on the number of handovers:
- If the number of handovers during the time period t-Evaluation exceeds n-CellChangeHigh
– the UE enters the High-mobility state
- If the number of handovers during the time period t-Evaluation exceeds n-CellChangeMedium but does not exceed n-CellChangeHigh
– the UE enters the Medium-mobility state
- else if criteria for either Medium- or High-mobility state is not detected during the time period t-HystNormal
– the UE enters the Normal-mobility state
Step 3: The UE scales the timeToTrigger (TTT) based on its mobility state:
- if the High-mobility state is detected:
– use the TTT value multiplied by sf-High
- else if the Medium-mobility state is detected:
– use the TTT value multiplied by sf-Medium
– no scaling is applied